G0~G8 성과·동맹 측정 OS 작업 일괄 고정
8월 7일까지 워킹트리에만 남아 있던 미커밋 작업을 커밋한다. 여러 사본 폴더(worktree·clone)에 흩어져 있던 중간 스냅샷을 정리하기 전에 원본을 git 이력으로 고정하는 것이 목적이다. - contracts/routes/services: measurement, outcome_trajectory, rupture_repair, deliberate_practice, calibration_transfer, supervision_research, multimodal_alliance, continuous_improvement 계열 신규 모듈과 테스트 - infra/db/init: 07~16 마이그레이션(측정 기반~calibration transfer 실행) - apps/web: 세션 리뷰 카드·관리 화면·E2E 스펙 추가 - docs/ops: G0~G8 라이브 통합·배포·롤백 증거 문서와 evidence JSON/PNG - scripts: smoke·ledger·릴리스 에이전트·NAS 프리뷰 운영 스크립트 engine.public 로그 .bak과 apps/web/test-results 산출물은 커밋에서 제외했다.
This commit is contained in:
Yun Chan
parent
93dd8f82d7
commit
16e791e044
390 changed files with 243188 additions and 499 deletions
173
apps/api/app/test_practice_runtime_observer.py
Normal file
173
apps/api/app/test_practice_runtime_observer.py
Normal file
|
|
@ -0,0 +1,173 @@
|
|||
from __future__ import annotations
|
||||
|
||||
import unittest
|
||||
from pathlib import Path
|
||||
from uuid import uuid4
|
||||
|
||||
from .services.deliberate_practice import (
|
||||
assess_practice_episode,
|
||||
load_practice_benchmark,
|
||||
prescribe_from_coaching_cards,
|
||||
)
|
||||
from .services.practice_runtime_observer import (
|
||||
EvaluatedTurnPair,
|
||||
RuntimePracticeObservationError,
|
||||
derive_runtime_episode,
|
||||
)
|
||||
|
||||
|
||||
BENCHMARK_PATH = (
|
||||
Path(__file__).resolve().parent
|
||||
/ "data"
|
||||
/ "deliberate_practice_benchmark_g4.v1.json"
|
||||
)
|
||||
|
||||
|
||||
class PracticeRuntimeObserverTests(unittest.TestCase):
|
||||
@classmethod
|
||||
def setUpClass(cls) -> None:
|
||||
cls.pack = load_practice_benchmark(BENCHMARK_PATH)
|
||||
|
||||
def test_durable_turn_pairs_produce_independent_before_after_episode(self) -> None:
|
||||
case = self.pack.cases[0]
|
||||
prescription = prescribe_from_coaching_cards(case.coaching_cards)[0]
|
||||
session_id = uuid4()
|
||||
same_case = uuid4()
|
||||
persona = uuid4()
|
||||
episode = derive_runtime_episode(
|
||||
prescription=prescription,
|
||||
practice_session_id=session_id,
|
||||
source_case_id=same_case,
|
||||
source_persona_id=persona,
|
||||
practice_case_id=same_case,
|
||||
practice_persona_id=persona,
|
||||
turn_pairs=(
|
||||
EvaluatedTurnPair(
|
||||
counselor_turn_id=uuid4(),
|
||||
counselor_turn_seq=1,
|
||||
client_turn_id=uuid4(),
|
||||
client_turn_seq=2,
|
||||
technique_codes=("facilitative_question",),
|
||||
client_state_codes=("defensive",),
|
||||
appropriateness="warn",
|
||||
utterance_fingerprint="sha256:first",
|
||||
),
|
||||
EvaluatedTurnPair(
|
||||
counselor_turn_id=uuid4(),
|
||||
counselor_turn_seq=3,
|
||||
client_turn_id=uuid4(),
|
||||
client_turn_seq=4,
|
||||
technique_codes=("reflection",),
|
||||
client_state_codes=("affect_contact",),
|
||||
appropriateness="pos",
|
||||
utterance_fingerprint="sha256:second",
|
||||
),
|
||||
),
|
||||
)
|
||||
|
||||
assessment = assess_practice_episode(prescription, episode)
|
||||
|
||||
self.assertEqual(assessment.comparison.change, "improved")
|
||||
self.assertEqual(assessment.progress, "transfer_pending")
|
||||
self.assertEqual(episode.attempts[-1].criterion.source_kind, "model_inferred")
|
||||
self.assertEqual(
|
||||
episode.attempts[-1].criterion.perspective,
|
||||
"independent_observer",
|
||||
)
|
||||
|
||||
def test_server_identity_marks_cross_case_as_unseen_transfer(self) -> None:
|
||||
case = self.pack.cases[0]
|
||||
prescription = prescribe_from_coaching_cards(case.coaching_cards)[0]
|
||||
episode = derive_runtime_episode(
|
||||
prescription=prescription,
|
||||
practice_session_id=uuid4(),
|
||||
source_case_id=uuid4(),
|
||||
source_persona_id=uuid4(),
|
||||
practice_case_id=uuid4(),
|
||||
practice_persona_id=uuid4(),
|
||||
turn_pairs=(
|
||||
EvaluatedTurnPair(
|
||||
counselor_turn_id=uuid4(),
|
||||
counselor_turn_seq=1,
|
||||
client_turn_id=uuid4(),
|
||||
client_turn_seq=2,
|
||||
technique_codes=("reflection",),
|
||||
client_state_codes=("thought_organizing",),
|
||||
appropriateness="pos",
|
||||
utterance_fingerprint="sha256:novel",
|
||||
),
|
||||
),
|
||||
)
|
||||
prior_state = case.graph.states[0].model_copy(
|
||||
update={
|
||||
"band": "consistent_local",
|
||||
"attempt_count": 1,
|
||||
"familiar_demonstrations": 1,
|
||||
"highest_familiar_difficulty": 1,
|
||||
"evidence_refs": tuple(
|
||||
case.episodes[0].attempts[0].criterion.evidence_refs
|
||||
),
|
||||
}
|
||||
)
|
||||
|
||||
assessment = assess_practice_episode(
|
||||
prescription,
|
||||
episode,
|
||||
prior_state=prior_state,
|
||||
)
|
||||
|
||||
self.assertEqual(episode.attempts[0].scenario_novelty, "unseen_transfer")
|
||||
self.assertEqual(assessment.progress, "mastered")
|
||||
|
||||
def test_unknown_competency_mapping_fails_closed(self) -> None:
|
||||
case = self.pack.cases[0]
|
||||
prescription = prescribe_from_coaching_cards(case.coaching_cards)[0].model_copy(
|
||||
update={"competency_id": "competency.unknown-skill"}
|
||||
)
|
||||
with self.assertRaisesRegex(RuntimePracticeObservationError, "unsupported"):
|
||||
derive_runtime_episode(
|
||||
prescription=prescription,
|
||||
practice_session_id=uuid4(),
|
||||
source_case_id=uuid4(),
|
||||
source_persona_id=uuid4(),
|
||||
practice_case_id=uuid4(),
|
||||
practice_persona_id=uuid4(),
|
||||
turn_pairs=(),
|
||||
)
|
||||
|
||||
def test_durable_audio_metadata_can_satisfy_voice_retry_without_capture(self) -> None:
|
||||
case = self.pack.cases[3]
|
||||
prescription = prescribe_from_coaching_cards(case.coaching_cards)[0]
|
||||
episode = derive_runtime_episode(
|
||||
prescription=prescription,
|
||||
practice_session_id=uuid4(),
|
||||
source_case_id=uuid4(),
|
||||
source_persona_id=uuid4(),
|
||||
practice_case_id=None,
|
||||
practice_persona_id=None,
|
||||
turn_pairs=(
|
||||
EvaluatedTurnPair(
|
||||
counselor_turn_id=uuid4(),
|
||||
counselor_turn_seq=1,
|
||||
client_turn_id=uuid4(),
|
||||
client_turn_seq=2,
|
||||
technique_codes=("holding",),
|
||||
client_state_codes=("thought_organizing",),
|
||||
appropriateness="pos",
|
||||
utterance_fingerprint="sha256:voice",
|
||||
has_voice_feature=True,
|
||||
),
|
||||
),
|
||||
)
|
||||
|
||||
assessment = assess_practice_episode(prescription, episode)
|
||||
|
||||
self.assertEqual(assessment.attempts[0].outcome, "passed")
|
||||
self.assertIn(
|
||||
"voice_feature",
|
||||
{item.kind for item in assessment.attempts[0].evidence_refs},
|
||||
)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
unittest.main()
|
||||
Loading…
Add table
Add a link
Reference in a new issue